Expedia × Profound | Competing in an AI-Driven Search World // With James Cadwallader and Daniel Shin Un Kang

Hey everyone, Jim here, and I've got some exciting news. The CMO Podcast is now on YouTube. You can now watch our conversations, not just listen. See the expressions, the energy, and the insights that happen when we sit down with the world's top marketing leaders. Head over to YouTube and subscribe to our channel, The CMO Podcast, so you never miss an episode and be part of the community. So please check it out, subscribe, and join the conversation.

Liberty Mutual. And Doug. There's nowhere I wouldn't go to help someone customize and save on car insurance with Liberty Mutual. Even if it means sitting front row at a comedy show.

Hey everyone, check out this guy and his bird.

What is this, your first date? Oh no, we help people customize and save on car insurance with Liberty Mutual together. We're married. Ah!

Me to a human, him to a bird.

Yeah, the bird looks out of your league anyways. Only pay for what you need at LibertyMutual.com.

Liberty, Liberty, Liberty, Liberty. What's the first brand you remember making an impact on you as a young boy? James, start with you.

I was big into soccer or football, as we would call it. Correct me.

I agree with that.

There were these David Beckham Adidas football boots. I can just remember it being so visceral. I needed to own that pair of football boots.

So Daniel, how about yourself?

I think the biggest brand is probably Apple. I think growing up in the middle of a farm, middle of nowhere, like getting access to stuff is hard. But then when I first came across like an iPhone and I said, wow, that's pretty incredible. What it can do. And it's had a thing for me since I was a kid. And, you know, I was finally able to use some Apple products. I'm very happy.

Hi, I'm Jim Stengel. I've helped hundreds of major brands discover and activate their purpose, because when a brand's purpose is clear, compelling, and authentic, profit naturally follows. Each week, I welcome the CMOs, the chief marketing officers of your favorite brands, to speak to how their job is so much more than marketing. These leaders share their inspiration and challenges along with how they try to build a full, healthy, and happy life in and out of the office. And it's that energy that reaches everyone they touch. And we're glad you're here to feel that energy and to learn from these remarkable leaders. So here we go. This is an especially important show today. We're living through the largest change in the internet since its existence, the shift from building content for people to building content for machines on behalf of people. In short, the rise of AI search, the shift in how we retrieve information about everything. My guests today are at the leading edge of that shift, and we will explore a case study on how to operate in this changing world with its implications for CMOs and frankly, all players in the marketing ecosystem. First, we have James Cadwallader, the co-founder and CEO of Profound, a fast-growing new company helping brands understand how they show up and compete in AI-driven discovery. James works with companies like US Bank, Chime, Expedia, and DocuSign in navigating the shift from traditional search to a world of answer engines, agents, and AI-led experiences. James founded Profound in 2024, has already raised about $60 million. His company has been recognized by Redpoint Ventures as one of the most promising private AI companies shaping the future of applied artificial intelligence. Joining James will be Daniel Shin Eun Kang, Head of Organic and Agentic Search at Expedia, one of the world's largest travel platforms serving millions of travelers globally and generating more than $12 billion in annual revenue. Daniel leads teams working at the intersection of search, AI, and customer experience inside a large, complex, performance-driven organization. Together, James and Daniel bring two powerful perspectives. This episode will really make you think. Let's get into it. A conversation about search, AI, and indeed the future of marketing. Well, James and Daniel, welcome to the CMO Podcast. I am super looking forward to this. We're recording this right before the Super Bowl. Now, Daniel, I know you're not super into sports, but I do have to say, will either one of you two have an ad in the Super Bowl this year?

I believe Expedia will have an ad, but let me, let me correct, make sure I'm not lying to the audience.

Right. Okay. He's looking it up. That's good. I suspect you will as well. I see Vrbo everywhere in sports. So how about you, James? Profound Super Bowl ad this year, maybe next year?

Yeah, I, you know, despite being a well-funded, uh, startup, I don't think we are at Super Bowl territory yet, uh, especially considering we are predominantly B2B. So yeah, no, I had no Super Bowl ad this year.

It's a matter of time, right, James?

A matter of time, yeah. Inevitable, but just not, not this year.

Okay guys, let's move from the Super Bowl and let's start with how you two met. I mean, we're going to talk an awful lot about this case study today of Expedia and Profound, but how long have you known each other?

So I knew of James far before James knew about me. So he's part of a community in San Francisco called South Harbor Commons, and when I joined, I think summer or so, we were talking at who are some of the top performers in this community, and James's name, of course, Profound, came up a couple of times. So he was on my radar for quite some time, but My story is I had joined Expedia originally as an entrepreneur in residence, and then a few months after became a P&L owner for the organic and agentic search. And as we were going through that as a practitioner, one of the obvious problems that comes up is how are we doing? And that's a surprisingly hard question to answer in terms of AI traffic and AI visibility. And of course, out of the 15, 20 companies that are working on the space, I talked to a lot of them. And James was already in conversation with Expedia at the time, and Profound is obviously the category winner. That's how we got to meet for the first time. I think our first real call was around contracting terms, and then after that we had a couple of other sessions.

Fun call. So Daniel, I'm going to ask you a really stupid, naive question. Why was that important for you? How are you doing Expedia in AI search?

Yeah, it sounds like an obvious question, but you have to know where you are in order to help get the bearings of where you want to go. And it's actually quite complex for a couple of reasons. The first is, unlike traditional search with a mature ecosystem of data and players, you don't actually get a lot of data services or ground truth. So I don't think anyone knows exactly what queries users are typing in when they're searching for brands like Expedia and otherwise. So what the industry has done, for the most part, is create synthetic prompts of what do we think users are going to ask these large LLM search engines. Scrape sort of the data and then get a proxy of where the industry is going. This is a really, really hard problem to do this at scale. You can imagine doing this dozens of times on a single platform. Now multiply that by, you know, 10, 12 different models that are out there and thousands of prompts, it becomes a nightmare. And this is why I think players like Profound exist, and that's why we got started with them.

Now I want you to talk about your first meeting was about a contract, and then you moved on to more fun stuff. So our audience is getting to know both of you, so I'd like I'd like you to talk a bit about each other. And maybe James, you start with— Talk about Daniel a bit. What inspires you about him? What strikes you as a person, as a leader about Daniel?

[Speaker:JAMES CADWALLADER] I mean, I think, well, I can speak to Expedia as a whole. So Expedia as an org are one of the most sophisticated teams in this category right now. I guess as quick, fast background, I'm James Cadwallader, co-founder, CEO of Profound. Yeah, we launched the business about 18 months ago in August of 2024. It's been a rip-roaring journey of, yeah, it's been exciting and fun and mayhem, all things together all at once. But on this journey, what's ended up happening is we now work with, I mean, and it's a true luxury, it's a true honor. We work with the biggest brands in the world, like quite literally, I think we, it's dozens of Fortune 500 companies. And you get very good exposure and start to understand which marketing teams are super sophisticated. Now, unfortunately, it makes it sound like I'm just sucking up to Daniel because I can only really talk about Expedia on this call because Fortune 500 companies don't want me sharing their, you know, who's doing what. But I can say that there's probably half a dozen that really spring to mind. And within that half a dozen, I'd say Expedia is probably in the top 1 or 2 in terms of their— how much the team is leaned into this category. How sophisticated they are. I went over to the Expedia HQ in Seattle and sat with the team as they spoke about AI search as a, as a category and left blown away by the level of sophistication, the level of granularity, the amount of content that is created and hosted on Expedia is once you really pay witness to it, you're like, whoa, this is a really big operation. And you can see a business that has been built on an incredible approach to marketing, one that I've never seen before. And now Daniel, within that, I think is unique in that he comes from a tech background. So Daniel comes from, you know, he doesn't really come from a, you know, your traditional marketing background, but he, he really has a kind of a hacker mindset and a very tech-forward mindset. I mean, that speaks to his time at South Park Commons, which for anyone who's not familiar is like a SF-based accelerator incubator program that was set up by two of the founding engineers of Facebook. It's pretty hard to get into. I mean, I say that as someone that was accepted, so it sounds a bit self-congratulating, but it's pretty hard to get into. And I think once I pulled two and two together and I was like, oh, wow, that's— yeah, Daniel from SPC, that's crazy. You're now at Expedia. That is a very unique offering, and it all kind of marries together and makes a ton of sense. So yeah, long answer, but that's my— yeah, that's my impression of Expedia and Daniel.

Yeah.

So Daniel, you knew about James before he knew about you. So now that you're working with him and you're getting to see the real thing, does he inspire you as much as he did before you knew him?

There's only one answer to that question.

I don't know about that.

Only one correct answer. Yes, that's right. I think for sure. So I'll talk a little bit about my background and it matters in this context of how I perceive James. So my background is I used to be a venture growth investor at SoftBank Vision Fund. So I invested in top growth equity companies for about 2 years, sat on boards, and sort of help them with their pre-IPO to IPO sort of phrase. And then I said, maybe I could build one of those companies, like any delusional founder. So I left that job, started my own company with a different accelerator called Y Combinator, and sort of scaled that company to some time. The reason I mentioned this is before you dive into, even as an investor looking inwards to founders, I don't think every investor quite understands what it's like to be a founder until you're a founder yourself. And having been in that seat, I have first tremendous amount of respect for founders who are actually building in this frontier technologies. They saw something that are non-obvious to other people at that time and then actually executed. Right. So I think that's really, really big. And I think Profound and James with his co-founders, actually one of the very first who was to really pioneer this, this industry. So I think that speaks to a lot. SBC, again, a little bit, a little bit of a sycophantic thing, but yes, I think it's a great program as well. But The way that James responds to problems, like in the early days, a founder is heavily focused on building what customers want and the rate of response, how quickly he responds and how thoughtful he is with all of these is just a testament. The fact that he actually flew over to Seattle, spent, I think, what was it, James, like 8, 9 hours, something crazy like this, where he was like locked in a room, just speaks to how much he cares about customers. And I would be very bullish on this company.

I appreciate that. I mean, one thing I'd add a bit of color to that, which is quite interesting. So we get to Seattle, it's myself and Kendall from my team who's worked super closely with the Expedia team. I don't know, I mean, maybe I was missing some context, but we were working with Alistair from your team, Daniel, and Alistair, he's a super smart guy, but like very nonchalant, like kind of brings us into the lobby, gives us a tour. And he's like, oh yeah, we're gonna go into the meeting room. But I thought it was gonna be like a few, you know, a few folks, sat around drinking a coffee or something and just super casually, we go walking around this corner and I walk into this room. No word of warning. No, like, just Alistair. I'm just following Alistair around the corner into this room. And I kid you not, there's about 25 people. It looked like one of those, like a president's briefing or something. It was like this giant horseshoe-shaped table arrangement with about 25 people all sat waiting for us to walk into the room. So it was a, it was a bit of a shock to the system that day. But yeah, it was about 8 or 9 hours in that room going through the details. Yeah.

Well, I know from my homework on you, James, you don't have a lot of hobbies. You love business, you love work. So this is all fitting together.

Hey everybody, this is Andrea Sullivan, the CEO of Vibe. And we produce the CMO Podcast and are so excited to have partnered with Jim Stengel for so many years. Wanted to tell you a little bit about something that could be right for you. VIBE is a program for entrepreneurs and business leaders who want to get more out of their life and become their best and happiest selves, both personally and professionally. We have a 12-month program that allows people to meet up with some of the best business leaders out there and additionally experts in the wellness sector so that you can learn how to nurture yourself. So we teach things around sleep optimization, meditation, all those good things as well. Please reach out to us at podcast@vive.co to get more information. That's podcast@vive.co. Thanks for listening to the CMO Podcast and hope to talk to you more about Vyve.

So listen, we're already kind of going here, but I think As context for the discussion about how you two are working together in this little case study of Profound and Expedia, we need to have more context on your company, James, and you're sort of already going there. It was founded in August 2024, not that long ago, but could you take us back to then? What was the insight? What was the problem you were trying to solve? Why the name you chose as Profound? Walk us through the garden of the history of your company a bit to give us context for what we're going to talk about in the work you're doing with Expedia.

Yeah, for sure. It's interesting. I think, I mean, particularly over the last 6 months, I've actually had a few people say to me in reference to the inception of Profound, like, wasn't it, it was a very obvious idea, or people will enter a conversation saying, it's a very obvious idea. But back in the beginning of 2024, it was so non-obvious. No one was, it was a very, very peculiar paradigm, because it was myself and my co-founder Dylan, who, he was a software engineer at Uber working on geospatial products. So lots of map building and building the rider app experience. And we were exploring lots of ideas. I was very interested in search, generally speaking, somewhat on the consumer side. I was interested just in information retrieval and what I think really Perplexity was pioneering in information retrieval was becoming more and more interesting. And really our business was kind of thinking of a second order effect where once you start to— if you've read books like The Search or Into the Plex or even Chaos Monkeys to an extent, and you've studied the origins of search or the origins of Google, which is kind of synonymous with search, the thing you begin to understand is that search or search engines, i.e., Google Search, is the biggest outcome really in the history of capitalism.

Yeah.

And that outcome is predicated on this idea of we will give you links and you will click them and then we will pepper in some paid links and you might click those as well. And now we have this platform shift, you know, with the birth of ChatGPT and generative AI, which counterpositions entirely against that idea where you no longer have to click anything. You can just talk to the internet and it talks back to you. So in that new world, what happens to the click? Like, it undermines the entire business model of Blue Link Search, which really is the biggest outcome in the history of capitalism. So once you start to realize that, it was this, like, I mean, it sounds a bit dramatic, but dare I say sublime, kind of like it was almost overwhelming at the time to realize it was like, wait, surely not. This can't be. Do you know that movie The Big Short?

Sure.

Do you know that feeling they capture so well where they're like rubbing their eyes? They're like, surely not. This can't be real. Why is no one else talking about this? That was the feeling as we were. Because, you know, you know how obvious this is now, this space. Like there's, you know, as you said, there's a new AEO tool that pops up every week and YC is backed like 10 in the space and it's a very, highly spoken about, extremely obvious category now. Imagine it being that obvious just to one person at the beginning when no one else was talking about it. So there's this big short moment in the very early days, and you really had to resist the urge to second guess yourself because it just felt like we had to be right. Surely not, because this is the biggest outcome in the history of capitalism. Surely not. And why is no one talking about this? But we just went all in. And our idea is that in a world where you can talk to the internet and it talks back to you, or when AI answers your question, the only thing that will matter to marketers is how AI talks about your brand, how AI talks about your product or services. And AI has to get that information from somewhere. It doesn't rely on the pre-training. It can't rely on the pre-training. So it has to pull from the open web, information on the open web. So now how you distribute content, ideas, information to bots on the internet that are crawling on behalf of AI systems like ChatGPT or Gemini or Grok or Meta becomes the most important priority in the history of marketing. So yeah, that was kind of a meandering, I don't know, glimpse into maybe how we were thinking back at the time.

So there's that provocative thought insight, rubbing your eyes, just can't— but then what did you do? I mean, you had to say, well, No one's talking about this. We should. This is a massive problem for every brand, right? So what role can we play in helping brands solve the problem? How'd you make the next leap?

I mean, it like, then we went to war. It felt like, you know, it was almost like you spent some time in the, you know, that with the everything prior to this point in the story had been like more, you know, the exploring the, the blue ocean thinking, the, the, the what could, what, what if. But once we crossed a certain threshold of conviction in the idea, it's like an entirely different mindset. It's you essentially put on your helmet, you grab your rifle, and you go into the trenches. And yeah, you know, immediately you just take— it's left foot, right foot, left foot, right foot. And it's just quick reps. It's okay, let's go, let's go and speak to some of the marketers that I know. Johnny Bentwood at Golem, got introduced to him before we had anything, we had a PDF, and he was super supportive. And there's all these kind of miracles, these mini miraculous moments along the journey, like as you'd hope and expect. Like, you know, we met Johnny Bentwood at Golan who didn't know us particularly well. I just got introduced to him through someone else that I worked with. But instantly he was like, that's a great idea. It's a great idea. We'll give you a $15,000 LOI, which to us at the time when it was just a PDF was like, that was a crazy opportunity. We were like, wow, there's someone willing to pay 15 grand immediately. And then it's like, okay, well, Johnny, what do you want? And he's like, oh, I want it to look like this and I want it to do that. And you just rep and rep and rep and rep. And my co-founder, luckily for me, I think Dylan is one of the best product engineers on the planet. He's very talented at what our type of business is. He's not, you know, he is not a scientist working at DeepMind on neural nets, but he is incredibly good at building product and he likes building product. He thinks product's fun. So once we started repping, it was, you know, then we found Charles, our founding engineer, who again, a miracle moment. He had written a dissertation essentially on this exact topic and then gone to Omnicom which is where he was working at Omnicom and he's bored out of his brain there.

And one of the biggest holding companies in the world in communications, just for our audience.

Yeah, I would say Omnicom is not known as a hub for the best engineers in the world. Put it that way.

I know that. Yeah, for sure.

It's not like hiring someone from DeepMind where it's kind of figured out and obvious. But Charles, our founding engineer, was at a funeral of a mutual friend with one of our mutual friends. And they were talking about our business and he was like, "I'm literally thinking about that right now." So then 3 days later, he jumped on a plane from Chicago and came down to New York and the rest is history there. So all these mini miracles on the journey. Yeah.

So good. So Daniel, I'm going to bring you in in a minute, but I want you, James, just for a moment to reflect on the last 18 months in your life. I mean, this has got to be the most important thing you've ever worked on, maybe that you will ever work on.

Yes.

Tell us a little bit about what these 18 months have been like for you as a human being, an individual, a person, a guy who loves business.

Yeah. I mean, I definitely identify as like a disciple of business. I find it interesting. My fiancée always laughs that, you know, if I'm relaxing on a Sunday afternoon, we work pretty hard here. I mean, in the office 6 days a week, and then Sundays, even when I'm at home, I'm quite often, you know, on my phone or doing stuff anyway. But my relaxation would be watching a podcast like yours, James, or maybe reading a great book on— I don't know, I'm trying to think like what are some of the aforementioned, listening to the Acquired series on Google. So I'm a disciple of business. And yeah, I think this is the most fun I've ever had. I don't want it to stop. And I, yeah, work. It's like the universe has given us this gift and the only way to reciprocate and pay it back is just to give it everything you've got. So yeah, I think, yeah, and you're right. I don't think I'm going to get another opportunity like this. I want to work on this until I die.

So Daniel, react to all that that James just said. What's buzzing around in your mind as he goes through that?

I think you call it a gift, and I think it's a humble way to put it. To your point very early on, James, it was obvious to you. And I think that's the benefit of being around a community that's forward-looking where non-obvious things become obvious, but there are very few people who actually act on that. So I don't know what path you took to arrive on that conviction, but once you got into conviction, you said, now it's war, right? And very few people are able to execute on that level of conviction because you literally bet your life on it when you had nothing, and your first win was a $15K LOI. And like, let's be honest, $15K LOI is not even a real contract. It's not a recognizable ARR metric. It's a document, right? But the fact that you're able to put left foot, right foot forward and just progress is fantastic. And yeah, I'm— and it goes back to the side story of how responsive James— he's a co-founder of this large org, but he's still yet extremely responsive because he's working. He cares about customers. He's looking to see where can we go next. And I think there's tons of products in the pipeline, James, I'm sure that you have that reflect a lot of the needs of the customers. So I'm, again, very, very Happy to have met a founder who is in this pioneer of this category and glad that we get to partner.

Well, Daniel, your last 18 to 24 months has been also interesting, right? You did a book, you launched a book, The Super Upside Factor. You joined Expedia in 2025, head of organic and agentic search. So talk a bit about your life the last 18 to 24 months. How's it been?

Yeah, it's been a little bit of a whirlwind and lot of learning experiences, I would say. So I started my company in mid-late 2021, and it was a, it was a company that wanted to allow financial services and access to as many people as possible, starting with freelancers and creators. And we were on a great trajectory and working on it for quite some time. But 3 or 4 years into the journey, I had one, gotten very tired, and two, there are many macro shocks into the business where it came, we came to a decision that it is not in the best interest of us, the investors or the customers to continue this company. And we still had to, it was a really decision because we still had, you know, over $2 million left in the bank. It was— we're at the biggest platform shift we've seen forever. I think everybody agrees this is probably one of the best times ever to do a startup. So walking away from that was quite hard, but we didn't think the motivations of not wanting to look like a failure or continuing just because we can was actually a very good decision for anyone involved in the party. So we made the decision to shut it down, return the capital. And I took some time off. And as that was happening, my book was released, which was really, really tough because on the face of it, I had to show, hey, I'm doing really well, it's successful. The book actually talks about applying venture math to careers. So the crux of the book is there's a trillion-dollar industry that's paid to be wrong. So venture capitalists lose money about half the deals they make, and they miss the mark 90% of the time. But if in the few times that they win, it's a 5,000x return, as we've seen with Masa's investment in Alibaba or indexes investment in Figma. I think Figma was a 6,000x return or something like this. You can still win dramatically even if you're wrong most of the time. And how do you sort of apply that into your career? And, you know, got to talk about this in the stages of TEDxStanford. But internally, as you can imagine, was I had just killed my baby of a startup in going through that. And I said, maybe I need to take a break from entrepreneurship for a little bit. But oddly enough, joined Expedia as an entrepreneur in residence, uh, and also got into look at the future of search a little bit more. Where is search evolving? So AEO was a big topic even when I got there and not formally involved. But I was thinking, all right, like, where's this thing going? What about the ads? What about the agent of commerce? What about all of these things? And I thought deeply about what I call probable future states. So rather than what will happen, what are the ranges of futures that would happen? And how would we, as Expedia or anybody, anybody else, position ourselves to these probable future states so that we're in an advantage. So as I was thinking through that, an opportunity came up where they said, hey, well, we really like your work. We think this is interesting. Can you go and actually run this P&L? So I think it's been, what, 3 and a half months now since I've been running this. So everything is very, very new. I'm getting tons of help. I'm learning a ton from, again, practitioners and operators at this frontier like James and like others who are building.

Now let's get into what you are working on together, what your companies are working on together. So, and I would love this as much as you can be transparent about it and open. To be a bit of a case study on James, what your company does to help brands, and Daniel, and what your company does to work with a team like Profound to understand what's going on in our ecosystem, especially around things like AI and AEO, or answer engine optimization, and leverage it to build your brands and grow your share and make earnings and have a great future for your company. So, I think we should start with what are you doing? What's the scope of your relationship right now? What's the scope of the work you're doing?

I'll kick off and set a little bit of the context of what visibility means and where we're going and where we're working together and where we want to— or I'll speak for myself— or I want to continue to work together. When it comes to AEO visibility, it's important for everybody to understand again where you are and what metrics you're going after. It's actually a surprisingly hard thing to crack, and I'll talk about at least 3 major complications. The first is, again, to set context, no ground truth. You have to synthetically prompt to get a proxy of what consumers are seeing on their screen and then try to get in front of them. So on the visibility front, there's a couple of hard problems you have to solve. The first is getting to the right set of prompts. How do you know that this thing is actually reflective of what consumers are searching? So how do you do that? The second is actually choosing a vendor and understanding their methodology. So Profound is one of them. There's like 15, 20+ companies that are doing this with different methodologies. An example of this is even in the prompting stage, some companies prompt using a headless browser, others use an API, and some have like some memory enabled. These small methodologies actually creates a large variance in the actual outcome that we see on the outside. So how do we resolve that? And what is the best methodology that best reflects what happens in the real world? And then the last thing you sort of have to think about is once you get this raw output, and by the way, this is like another technical hard thing you have to do, but once you get that raw output, you then need to derive a visibility metric. So when people say visibility, what does that even mean? Is that like a rate? Is that a share of voice? Like, what does that mean?

And when you say visibility, it's your brands, right?

That's correct. And how is this actually correlated with AI traffic? Right. So at the end of the day, yes, brand exposure is important. All of these is important. But for a consumer brand like ours, getting that revenue and getting that transaction and making sure it correlates is important. Otherwise you're optimizing for the wrong variable. So there are like 3 really hard problems to solve. And I think internally we've done a great job of this with the help of our partners. And where we're working together very actively is the synthetic prompting and the vendor methodology. I think Profound has done a good job of understanding the trade-offs and coming up with methodologies and also giving guidance on how to think about even the prompting and scraping of the information. I'll give you like one very concrete example. When prompting, some vendors choose to prompt via the API, and there's a lot of business benefits to that. It's, it's cheaper, it's scalable, it often breaks less. You probably need less stuff to scrape the data, but I don't think it's as reflective of what consumers actually go through. So despite the fact that it's more expensive, that there's a lot more operational overhead, you know, Profound has chosen to take this methodology for that accuracy, right? So these are like the snippets of examples of why, you know, working with a partner like Profound is helpful.

For a non-technical person, what is the exact business problem that you are trying to solve? And you say it's visibility of your brands, but take that a step further.

Okay. So basically you want to maximize in a new platform shift to maybe relate it to James's view. Google used to show a bunch of links and route people to different pieces of information. AI search engines ingest information and then share the most relevant, how do you show up as a part of that? And whether it's in brand mentions or links. So you want to know how much, where you are showing up and you're not showing up, and then constantly optimize your stuff to be able to go do that. And I can expand a little bit further of how AEO is a simple but a hard path to success. Essentially, you want to get into one of four inputs. The first input is the training data, and that alone isn't enough, but you do want to be included in somehow. The second is a retrieval process. So when a user types in the queries, it looks at the model, but sometimes it gets snippets of information from the web and then surfaces it. The third is outside your direct control, but you want customers to actively query your brand into their queries, and then you show up in your context window. And the last thing is, as these things compound, whether users mention your brand in their query or the output contains your brand, this gets consolidated into your user memory and will be referred to. And if you can maximize that, you get a compounding effect of your brand increasing, right? So you want to be included in one of those 4 inputs. The way you do that is through 2 things. The first is technical enhancement. So is our content even discoverable is like a big one. And then the second is worthiness. Is your content worthy of showing to our end customer? So when you look at those 2 things, there's a ton that you could do, everything from file types like markdowns to make sure that bots you know, are able to access your stuff? And it turns out different bots react differently to different file types, right? Or do you even recognize what bots are crawling on your website? All this technical stuff. And then the content is, of course, you know, is it relevant? Is it trustworthy? Is it actually useful? And, um, to, to the end user. So it's simple in a sense that there's 4 inputs and you have 2 levers to get there, but getting there is actually pretty hard. And visibility, of course, helps you understand where you're missing, what is working and what isn't working. When you're trying to optimize these two levers.

James, I want you to talk a bit about how you're helping with those business problems, but I love a quote that, as I was looking at video of you, James, before we started, and this relates to what Daniel just said, "We're not building content anymore for human beings. We're building content for user agents on behalf of consumers or human beings. That's such a profound shift. And Daniel, I think everything you talked about was about that shift. And at the end of the day, you have to have things that are built for the bots, but that people really need and want. And that must be very, very tricky to manage. So James, could you talk a bit about your company in terms of how you're helping Daniel solve that very enormous opportunity/problem?

Yeah, I think the point you make is a good one in that it's just true that you are creating— I said marketing to machines. Really, that's the idea, right? Because if I ask ChatGPT a question about where's a good place to go on vacation for my honeymoon in the Bahamas, where does ChatGPT go to answer that question? Does it use the pre-training that Daniel mentioned, or does it trigger a web crawl? Because I say, give me the latest hotels, or find me something very specific that the pre-training doesn't feel, you know, the classifier would say the pre-training probably can't answer this. So then, okay, well, does it? It's going to run a bunch of web searches and it will surface likely some Expedia results. But then how does that content get pulled through? Because the content, a lot of the content was designed for humans. Now, I won't overshare out of turn, Daniel, just in case I'm making you nervous on the other side of the screen. I won't share any secrets. You guys have a lot of sophistication around what you do. But I mean, I think one thing that is kind of obvious about Expedia is they just have huge volume. Daniel, can you speak to any of the volume that you have in terms of web? Because that was one of the most fascinating things that I discovered through our partnership is just the sheer volume of content these guys have is just hard to wrap your head around sometimes.

Exactly. And picking a partner that's able to work with that scale was hard. We're talking hundreds of millions of assets, right? So try optimizing that. I'm sure James, you'll talk to this, but You're going to need a machine to be able to do this.

Yeah, I mean, that's— so it's enormous quantities of content. And I think for what it's worth, there's two ways to interpret that. You could just interpret it as that, you know, Expedia is N of 1 and content is their business. So you'd expect them to have hundreds of millions of pieces of content. I also think there's— and it sounds like sycophancy, but I think there's an argument just to say that they are a very sophisticated team and that they are pioneering the future of content marketing. And actually, most businesses will catch up, maybe not to that same order of magnitude. But like, I think businesses are going to create a lot more content, put it that way. And you need great technology to understand how that content is performing at scale in these new AI surfaces. So if we ask, you know, 100,000 or a million questions about vacations in the Bahamas, how does Expedia get pulled through both as a brand and as it's mentioned, but also does it get cited? Is it going, you know, are the models using Expedia.com or, you know, JamesVacationBlog.com to answer these questions? And that varies from platform to platform. So really giving the team visibility into not just how they show up, but also the why, you know, why, why do you show up here? Why don't you show up here? And then A place that I think we— I mean, candidly, without, you know, I don't think this is super confidential or anything, but I think a place that we're excited to lean in more with the Expedia team is on sentiment as well, which is kind of like a new vector. Sentiment didn't really exist in the world of traditional search because as we've spoken about, Google Search was just surfacing links. There's no sentiment attributed to a link. But, you know, think about it like this. If I ask ChatGPT, what are the best marketing podcasts in the world? It will say something about the CMO Podcast, but it will give an opinion as well. You say Jim always gives really, you know, Jim, Jim always gives hot takes about XYZ, but sometimes it misses the mark on this and this. So you're getting an opinion for free. And that opinion varies and evolves over time when you're running hundreds of thousands of prompts. So, you know, every day, you start to see the drift of that sentiment starts to change over time, and then you can attribute why that sentiment is changing. So it's a new surface area. I think really what this comes down to over time, though, which is going to be— when I was watching, I was actually watching a fantastic podcast with Winston, who's the founder of Harvey AI, a completely different category of business to ours. But he was with Brian Halligan, who's the founder of HubSpot, as a new podcast, and Winston was on that pod. And he had this amazing quote where he said, you have to re-earn your right to play in the category every 6 months. That's like the new world of building software. And so the next 6 months for us are going to see a huge focus on profound winning content orchestration as well. Because what we see is that, you know, as Daniel mentioned, all right, cool, you want to update 200,000 pages. Yeah.

Good luck.

But now because of generative AI, we can, and we can do it thoughtfully. It doesn't have to just be, I think the headwind here or the controversy is that, oh, that means AI slop. I don't think that's true. Slop is a function of context. If you give models tons of context, the output's amazing. Slop is, consequence of if I just said to ChatGPT right now, create 10,000 blog posts for Expedia.com, it would happily do it, but it would suck. But if I gave it tons of context and raw data and interesting inputs and unique insights, it does a very good job, or at least it can get it 90% of the way. So yeah, I think Profound will become this workbench that different teams at Expedia use to orchestrate and automate production of quality content. And it will start off as quite low lift, kind of, you know, the simple content. It's not necessarily going to be aimed at humans. But yeah, that's where we see things going.

Daniel, what's James's team helping you do that you could not do yourselves?

I think there's a central point of we oversimplify it by saying, hey, synthetic prompting. I think just get us the data. Here's the queries. And from a buyer's perspective, it seems like a simple ask. If you try to build it yourself, I'm sure it's technically very, very challenging to do, especially QA, making sure all the data pulls through, making everything, and making sure everything's accurate. And to James's point, also do then start doing pattern recognition, right? Of every search engine actually rewards different things, whether it's AI mode or Gemini or ChatGPT, Claude, they value different things. The sentiment on it is going to be different. All of these things are something that is, I think, probably very distracting to a company that is doing something else. In our case, meeting travelers where they are. We don't want to spend all of our time doing this. This is something we need to help diagnose the problem and start optimizing our content and our technicalities. And this is where Profound is helping. And to James's point, I think as we go forward and we integrate further, it would be great to sort of— how do we think about content production and QA? Of course, we have our own engines, but I think there are different areas where we can play together.

I think this, I mean, there's a heuristic to that as well, and it is a heuristic, but I think you could also just take the simple point that we've got close to 50 engineers or folks working in engineering, product design, data now, working full-time on this. And by full-time, I mean mostly 6 days a week, long days, long nights, and have been working, have been working on this for a year and a half, day in, day out, day in, day out. And the product that we give to Expedia is the output of all that work.

Do you think we'll ever get to a point where there will be a way for us to assess how we show up as a brand in the AI search world? You talk about sentiment, we're starting to understand that, but if I'm one of Expedia's brands, if I'm Coca-Cola, if I'm Toyota, I want to show up right and I want to have an idea to Am I behind, ahead? Where do you think that's going to go?

Yeah, I mean, we essentially, that's what we do today, honestly. We work with the biggest brands in the world to help them understand how they show up on an eval-based approach. But also we have a very sophisticated suite of CDN integrations. So we can actually plug in as well to, if it's Akamai or Cloudflare or AWS, wherever the site is hosted, And then we're able to even see when ChatGPT is visiting that page to cite the content. So we can actually give you that. That's first-party data. But most marketers aren't using that data because it's, you know, it sits with IT or, you know, web services. But we expose those user agent visits to marketers. And essentially that's the new click. When a user agent visits the cmopodcast.com, that's the replacement for a human being because it's crawling on my behalf. So yeah, I mean, essentially, long way to say, I genuinely believe we can give you at least a good direction of understanding there today.

And I 100% agree with that. And I think what we're doing is, it sounds a little cringe, but it is world-class in terms of the industry standard. But I'll talk about some of the limitations there, but things like, hey, understanding user agent string, like who even visited you, and then working through, again, the, the attributions and the what the data actually means. So a simple example might be like, hey, out of all of the results, like where did you rank in results? As in when they say, hey, top 3 destinations, were we, you know, on the second one or the third one? All of these things. And you can start fine-tuning to your point, Jim, like where do you show up and what does that mean? But I think what we have is extremely helpful for setting a direction of how we're doing and where we're going. Of course, I think there are some industry problems to be solved. I think James and I talk about this from time to time, and Profile Art is a great solution for it, building towards this. But user memory is going to be a problem, right? So as we think about how the construct works, which is we are prompting search engines as if users were doing them, looking at the output, and then using that as a proxy. But as more and more users use AI search engines, their memory and personalization will come into effect and impact what the end outcome will look like. So if you think about the RAG process, retrieval augmentation and generative, more and more the personalization, the context window and memory is going to impact it. So how, as this scales over time, what is the accuracy of this proxy? And I think that's an industry problem to be solved. And James, I think you're already working on a few solutions, right?

Yeah, we have persona-based prompting, so we can build personas. So we could build a persona of people between the ages of 30 to 35 who are based in Dallas with an interest in horseback riding and Mexican food. We can then send prompts through that persona. And I think, you know, I definitely, I'm always the first to kind of like nod my hat to Daniel's point. It's not surefire, it's not perfect, but it starts to, you definitely see once you append that persona, to 1,000 prompts and you send the prompt through that persona, you see a difference in the results. So it's definitely true that you can start to see that difference.

We're living through the biggest change in the internet since we've known it, right? And you talked about that, James, about a half hour ago, and Daniel, you've been talking about this. What's your counsel? Both of you, because you're deeply into this. You're doing super fascinating work. We have a lot of CMOs in our audience who are scratching their head right now saying, "Oh, am I doing enough? What should I be doing? Am I preparing myself? What are we doing that's needed to win in this new world?" So what's your counsel to someone who's watching or listening who feels an urgency to get going? What should they do tomorrow?

I'll zoom in a couple of things. I'll make two points. The first is we talked a lot about AEO today, and it's certainly really, really important. And for most companies, it could be the vast majority of your AI traffic. But I think thinking of AEO in silo with the rest of Agentic Search is a mistake. So AI ads is coming. Agentic Commerce is already here. Google launched their UCP, the Universal Commerce Protocol. OpenAI has their ACP. These are all words, but like the idea of finding product market fit in each of these surfaces and platforms is going to be extraordinarily important for every consumer brand that's out there. I think if you can't get AEO, like, good luck getting everything else. But it's— you have to think of it as a portfolio. And the search engines will also think of them as a portfolio of where to send the users on a query basis as they compete with other search engines to retain user attention. So I think that's point number 1. Point number 2 is a little bit more theoretical and zooming out a little bit more. Every major tech company, I think, is playing AI bingo. And what I mean by that is they're not solving for a single future. They're solving for many futures. And I talked about these probable future states. For example, Meta acquired Manus late last year. Right now, Manus is a general purpose agent that allows web rails to complete tasks on behalf of people. OpenAI launched Atlas to compete with Comet. All of these things are not isolated products, but I think a coverage play. So I'll give you like a very quick mental model to think about it. Imagine on one axis you've got the human interaction point, where humans interact with agents. And then another axis, you've got agent rails, how agents communicate with one another. So right now on the human side, there's a lot of assumption that the primary interface is apps like ChatGPT and Claude and Perplexity. But it's possible it happens at the browser level. It's possible happens at the OS level, like your phone and your computer and your hardware level, right? And then on how agents communicate, there's MCPs, which seems to be the primary driver today, but there's UCPs and ACPs and also a host of payment regulation problems that need to be solved for agentic commerce to scale. So if you think of this and you see what tech companies are doing, they are precisely playing some version of bingo. And I think at any tech company or any CMO, you should, yes, weigh in on what's happening today, but what will happen and what will be the dominant form of where consumers interact with AI in the future and make sure that you're positioned at an advantage for whatever future that is. So that's my two cents.

James Cadwallader.

Yeah, my counsel to any CMO right now would be to go and find whoever was running your SEO, give them a pay rise, and give them a pay rise and realize that what was formerly considered the kind of black sheep of the marketing stack could well now become the most important priority across the entire business. There's an ongoing debate around, you know, SEO versus AEO, GEO. All I'm saying is those folks that have been used to building content for algorithms are very well positioned to thinking about this new problem space of building content that's designed to be consumed by bots. And I think they are best positioned with this kind of— they have all the toolsets of kind of a technical know-how, technical ability combined with— it does still involve some creativity. So I think that SEO function could become mega sexy is basically my view. And if you have a good SEO team in place, both internal and agency, I would lean in. Maybe to match Daniel's double point, the second thing, it's not a kind of one-and-done big red button. This is a function of marketing. You know, how you show up on these new surfaces is not— you don't solve it. It's an ongoing endeavor. And you need to give— if you empower that team, you need to give them the space and the encouragement and the, you know, the necessary resources to be able to build that strategy. We work with incredible marketing teams, but it's folks like Daniel or like the team at Ramp who are just super empowered and they are able to ship and iterate and learn with lots of agency. So yeah, that'd be my two cents.

Before we jump into the creator brief, let's talk about advertising for a minute. What do you think is going to be the biggest change in how we create, place ads in this world of creating content for machines on behalf of people? I mean, that seems to me to be a very significant change. What's your counsel on that?

My personal take is that generative ads will be the biggest unlock in marketing history. You know, if you think about the progression of advertising, beginning with newspapers, going to radio, television, social, it's essentially been a, I mean, search on the curve of intent, being able to measure intent, being able to understand what the user wants. You essentially, back in the days of newspapers, you were confined to just, okay, well, they bought the Financial Times, they probably work in finance. That was the best stab in in the dark. Google Search gave us a bit more. Social gave us interesting new signals because you could see what videos and comments, etc. Generative chat, multi-turn conversational interface, you understand such rich context and intent. Advertising is going to be so wonderful. And LLMs are perfectly suited to be able to synthesize a perfectly tweaked, tailored advertising moment exactly to that part of the conversation when Jim was asking about the Bahamas and just mentioned that, oh yeah, your partner loves paddleboarding. What if you booked this spot around the corner while you were there?

Personalization in a way that we could not imagine 3 years ago.

I'll share my speculation on the AI ad space a little bit. I'm a little bit torn because there's a lot of problems to be solved precisely because of, I think, what James mentioned. I think it's the curse of context. The amount of noise in conversations to find out what the true intent is, is going to be extraordinarily challenging. And then the bar for recommendations would be really high because you expect to be so personalized. So that's going to be a really hard problem for these companies to think about solving. How do you take a slob of text over the last 3 to 5 minutes of just text with disjointed ideas and get to the intent such that the correct ad is sort of placed? And I think there's going to be a balance in terms of how these search engines also think about placing ads compared to organic, compared to actually having an agent complete a task through Agentic Commerce. Right. And it's going to be an interesting one because they're all fighting for the users. And right now it's— yeah, it's going to be an interesting one for these new players to actually take dominance. I'll take like one other example of this is we know that AI search engines stole queries from Google. A lot of the queries are informational in nature, not commercial in intent, right? And it's because they can't service commercial. So it will take time for these AI search engines to catch up. But once they do, it'll be, uh, it'll be an interesting battle to see how the ads play out. What a lot of startups are doing right now is there are now third-party AI ad sort of services that are being launched, and there are publishers, um, in AI search and chat beyond the majors like Claude and ChatGPT that are actually sizable in their, their user base. It'll be interesting to see how this ecosystem develops instead of like having a couple of big players competing. Maybe there'll be this fat long tail that are competing alongside them. So it's an interesting one from a marketer's perspective. I think it'll be great to have a diversified sort of set of paid ads platforms, certainly. And then if there's obviously this long tail, even better for being specific on consumers.

So very optimistic you both are about the future of advertising.

Yes. Probably biased.

Okay. Okay, listen, let's flip into a very quick creator brief. You guys are both two amazing young men. And what I think is interesting is Daniel, you are super educated, right? Degrees, McGill, Oxford, others, you're an international pilot. James, you have no formal education. So I'd like you to both speak to that. Why was that the right choice for you?

I mean, who's doing better? I mean, I think, I think James is crushing it.

I remember at SPC when someone said, this is SPC, that incubator accelerator that I spoke through. They start you off in a cohort, so you're in a big group. And I remember I walked in one day, one of them was like, oh yeah, you're the uneducated one. I was like, damn, that's a brutal way to put it. Yeah.

But this was a good choice for you, James. Why?

I started my first business when I was 18 and just never looked back. So there was always— I mean, not now, but I think when I was 18, the idea was that I would start a business and then I was going to go back and do university. And then that didn't happen. Then I started a second business after that. It would be disingenuous of me to sort of make it sound like a Machiavellian strategy that was in place. I think I love learning. Quite often, especially for all the stick that it gets, I think the US education system is actually pretty cool. I think Stanford is a very exciting place. Same probably for Harvard or Oxford, I think, in the UK where Daniel was. And yeah, sometimes I fantasize about what it would've been like to have had a formal education, but I didn't. And here I am at war with my helmet on and my bayonet, and that's that.

Daniel, how about yourself? Quite an interesting academic and successful academic record. Why was this right for you?

I think, similar to James, I don't think there's this grand plan, but I think it had to do with my survival strategy. So I grew up in a low-income immigrant family in the middle of nowhere in Canada. So opportunities were hard to get by, and if I didn't make it, the downside was pretty catastrophic. So what are some ways to hedge my risk and elevate my base case every time I go somewhere? So when I trained to be a pilot, you know, I couldn't go ahead for a number of reasons, and then I pivoted to college. So college set a baseline of I could probably get a minimum wage job, hopefully. And then Oxford did the same thing. And my career throughout, whether it's SoftBank or starting my own company, sort of did that same thing. I think what I'm coming to realize is I've elevated my base to a level that I'm comfortable. And in the same way that maybe James fantasizes about the counterfactuals of what it would have been like, I think starting early and doing what pulls you gives you a higher chance of finding and arriving at your life's work. Right? Like, what do you want to dedicate? James, you said you're going to work on this for the rest of your life. Like, that's life's work. And I think that's incredible that you found it. And it's something that I will be looking forward to and continuing to iterate and seeing. I've built this base for myself. What can I go do with it and bet my life on it?

You know, this question's coming if you listen to our show. What's the first brand you remember making an impact on you as a young boy? James, start with you.

I was big into soccer or football. We would call it correctly.

Uh, I agree with that.

There were— I mean, there's some controversy right now, I think, with his son Brooklyn going around on— my fiancée was showing me it last night on social media. But there were these David Beckham Adidas football boots that were just like, uh, had to have them. Yeah. I mean, the only commentary I'd add is that I can just remember it being so visceral. Like, it wasn't— it was a needed-to-own that pair of football boots. Nothing else mattered in my world at the time. And I think it's interesting when I think we— actually, I shouldn't say who we work with and who we don't, but we may or may not work with the aforementioned brand in some way at this point in time. And it definitely leaves a lasting impact.

Yeah, sure does. I'm older than you guys by a lot, but I remember when Adidas first came into the US, basketball shoes, and it was that, and they were triple the price of what I was wearing, but I had to have them. And I really did feel like I could jump higher, move faster, perform better with them on.

I'm sure you did. I'm sure.

I'm sure. 90% of it, right?

Yeah.

So Daniel, how about yourself?

I think the biggest brand is probably Apple. I think growing up, again, in the middle of a farm, middle of nowhere, like getting access to stuff is hard. But then when I, when I first came across like an iPhone and I said, wow, that's, that's pretty incredible, I wonder what it can do. And it took some time before I got my first set of Apple stuff. Now everything's Apple in my, my, uh, desktop here. But, uh, I like that there are technologies and software specific to Apple that enable to you. Like, I do pop music, right? Songwriting is my hobby, and there's specific software that is just better on the Mac, and sound and plugins that are just better. So it's had a thing for me since I was a kid, and, you know, I'm finally able to use some Apple products. I'm very happy.

Jim, can I ask you a question? What AI model are you using most right now?

I'm using ChatGPT Pro. I'm using Gemini I want to start using Claude. I think Anthropic's an amazing organization, so I want to lean in, but I haven't done it yet. But I mean, for my needs, which are more research-oriented, I mean, I was being honored at a big thing in New York in April, and I just said, what should I wear? I had the most amazing response. I mean, options. They talked about my personality, the place I'm going, what it's like, I asked it the other day what I could learn from Acquired to be a better show. Amazing response. So a thought partner and a challenger, and I would say a standard elevator in addition to the normal stuff. But yeah, it's a different— I can't search and get that. So no, it's a game change. But I think I have a personal goal to expand and try some other models and just keep exploring. What do you recommend for what I do?

Why do you use ChatGPT and Gemini?

Gemini, I'm a Google shop, right? We do everything on Google, so it's just obviously easy. And ChatGPT, I guess, was the first that came across and I just signed up right away. Yeah, it's been life-changing.

I mean, I was ChatGPT pilled to the max. I was an early adopter and I think being an entrepreneur, you kind of root for the underdog like David and Goliath. I hate to say it, but more and more frequently as of late, I've been using Gemini because the responses are pretty good. So that's kind of—

no, they are good and it's easy. It's built into everything I do, right?

Yeah. So it's a lot for grabs still.

Yeah, no, for sure, for sure. So, okay guys, I think we have to end with this question. What should— I think you've kind of answered it, but I want you to put a punctuation mark on it— what should CMOs, senior marketing leaders do after they stop watching and listening to the show?

www.tryprofound.com.

I think you'll get a few of those.

I think they should, you know, I think they should pay attention to, to the category. I think you should understand that this is a very big platform shift. This is not a new— this isn't search with a new UI. You are building marketing for machines. This is the biggest platform shift in the history of marketing. And you should think about it and, you know, consider it and invest accordingly, is my opinion.

Yep. Daniel, anything to add to that?

Just from a practical point of view, as a CMO, this is one portfolio of many, you're probably juggling paid loyalty, like everything on the stack. Realistically, keep open a radar, but I get someone on your team, empower someone on your team to go deep on this stuff, and coordinate for you and bring the results back. Like you will not be coordinating with all of the paid channels and the marketing channels and the corp dev and everybody internally within your company to go and execute whether you build a product internally or work with a vendor. Like, you will not have the bandwidth. Go empower your team to go do that and just let them run.

Amen.

Love it, guys. Thank you. Did you enjoy this?

I loved it.

Yeah, it was great.

Let me add one thing to Daniel said. I was watching the college football finals this week, you know, Indiana against Miami. Apple, I think, released for the first time an ad about students and why they're right for students of all kinds. It is a gorgeous beautiful, thought-provoking ad. So if you haven't seen it, it's a long one. It was like 60 or 90 seconds. Lovely. Absolutely lovely.

Yeah, we'll take a look. James, thanks for thinking of being with this podcast. Jim, thanks for hosting me. This was great.

Yeah.

Really appreciate you having us on, Jim. It's truly— I'm not saying this to be polite— truly an honor and a privilege to be on with you. Thank you.

That was my conversation with James and Daniel. Listen, I normally do 3 takeaways for your This is Brandon, like on this one, but this discussion was so interesting. I think there is really just one takeaway for you, our viewers and listeners, and that is learn about this profound shift in the internet and the implications on marketing. I think James Cadwallader said it well. Go to the people who run search for you and get to know how you're going to show up in this new world of AI and creating content for machines on behalf of people. So go learn about it, talk about it with your team, elevate this on your priorities. This is really, really powerful. The implications for your brand and this industry are really deep. That's it for this week's episode of the CMO Podcast. As always, I would be grateful if you shared our show with your friends, along with subscribing and leaving a review on Apple Podcasts, Podcast, Spotify, or wherever you listen. The CMO Podcast is a Vybe original production. The views and opinions expressed by podcast speakers and guests are solely their own and do not reflect the opinions of our sponsors or its personnel, nor do our sponsors advocate or endorse any individuals or entities featured on the episodes.